About dataset : https://www.ons.gov.uk/census/census2021dictionary/variablesbytopic/healthdisabilityandunpaidcarevariablescensus2021/generalhealth

1. Data Collection¶

In [1]:
# Importing required dataset

import numpy as np # linear algebra
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
import seaborn as sns
import plotly.express as px # visualization

%matplotlib inline
import matplotlib.pyplot as plt

import warnings
warnings.filterwarnings("ignore")
In [2]:
# Loading UK 2021 census dataset

df_census_2021 = pd.read_csv("Dataset/Ethnic_census_data/2021_data/data_census_2021.csv", index_col=0)
df_census_2021.head(2)
Out[2]:
LSOA LAD Output_Areas Asian Black White Others
0 E01000001 City of London E00000001 5 3 155 13
1 E01000001 City of London E00000003 24 5 209 17
In [3]:
#Removing the column geography form LDA dataset before joining LDA and LSOA dataset

df_census_2021.drop(df_census_2021.loc[:,['Output_Areas']], inplace=True, axis=1)
In [4]:
# Selecting only Bradford for further analysis

df_census_2021 = df_census_2021[(df_census_2021.LAD == "Bradford")]
In [5]:
# counting unique values in join_data

n = len(pd.unique(df_census_2021["LSOA"]))
print("No.of.unique values :",n)

No.of.unique values : 312
In [6]:
# Checking the shape of the file

df_census_2021.shape
Out[6]:
(1575, 6)
In [7]:
# Loading the health dataset

census2021_ts037_lsoa = pd.read_csv("Dataset/Health/census2021-ts037-lsoa.csv", index_col=0)

census2021_ts038_lsoa = pd.read_csv("Dataset/Health/census2021-ts038-lsoa.csv", index_col=0)

census2021_ts039_lsoa = pd.read_csv("Dataset/Health/census2021-ts039-lsoa.csv", index_col=0)

census2021_ts040_lsoa = pd.read_csv("Dataset/Health/census2021-ts040-lsoa.csv", index_col=0)
In [8]:
#Removing the column geography form LDA dataset before joining LDA and LSOA dataset

census2021_ts037_lsoa.drop(census2021_ts037_lsoa.loc[:,['geography']], inplace=True, axis=1)

census2021_ts038_lsoa.drop(census2021_ts038_lsoa.loc[:,['geography']], inplace=True, axis=1)

census2021_ts039_lsoa.drop(census2021_ts039_lsoa.loc[:,['geography']], inplace=True, axis=1)

census2021_ts040_lsoa.drop(census2021_ts040_lsoa.loc[:,['geography']], inplace=True, axis=1)
In [9]:
census2021_ts037_lsoa.shape
Out[9]:
(35672, 7)
In [10]:
census2021_ts037_lsoa.head(1)
Out[10]:
geography code General health: Total: All usual residents General health: Very good health General health: Good health General health: Fair health General health: Bad health General health: Very bad health
date
2021 E01000001 1475 859 468 119 18 11
In [11]:
census2021_ts038_lsoa.shape
Out[11]:
(35672, 8)
In [12]:
census2021_ts038_lsoa.head(1)
Out[12]:
geography code Disability: Total: All usual residents Disability: Disabled under the Equality Act Disability: Disabled under the Equality Act: Day-to-day activities limited a lot Disability: Disabled under the Equality Act: Day-to-day activities limited a little Disability: Not disabled under the Equality Act Disability: Not disabled under the Equality Act: Has long term physical or mental health condition but day-to-day activities are not limited Disability: Not disabled under the Equality Act: No long term physical or mental health conditions
date
2021 E01000001 1475 152 36 116 1323 140 1183
In [13]:
census2021_ts039_lsoa.shape
Out[13]:
(35672, 10)
In [14]:
census2021_ts039_lsoa.head(1)
Out[14]:
geography code Provision of unpaid care: Total: All usual residents aged 5 and over Provision of unpaid care: Provides no unpaid care Provision of unpaid care: Provides 19 hours or less unpaid care a week Provision of unpaid care: Provides 9 hours or less unpaid care a week Provision of unpaid care: Provides 10 to 19 hours unpaid care a week Provision of unpaid care: Provides 20 to 49 hours unpaid care a week Provision of unpaid care: Provides 20 to 34 hours unpaid care a week Provision of unpaid care: Provides 35 to 49 hours unpaid care a week Provision of unpaid care: Provides 50 or more hours unpaid care a week
date
2021 E01000001 1422 1322 67 52 15 18 9 9 15
In [15]:
census2021_ts040_lsoa.shape
Out[15]:
(35672, 5)
In [16]:
census2021_ts040_lsoa.head(1)
Out[16]:
geography code Number of disabled people in household: Total: All households Number of disabled people in household: No people disabled under the Equality Act in household Number of disabled people in household: 1 person disabled under the Equality Act in household Number of disabled people in household: 2 or more people disabled under the Equality Act in household
date
2021 E01000001 838 701 123 14

2. Data Preparation¶

In [17]:
#Joining 2021 health data

join_data = pd.merge(left = census2021_ts037_lsoa, right = census2021_ts038_lsoa, how = 'left', on = 'geography code' )
In [18]:
#Joining 2021 health data

join_data = pd.merge(left = join_data, right = census2021_ts039_lsoa, how = 'left', on = 'geography code' )
In [19]:
#Joining 2021 health data

join_data = pd.merge(left = join_data, right = census2021_ts040_lsoa, how = 'left', on = 'geography code' )
In [20]:
# Renaming the column in 2021 census data for further joining of data

df_census_2021 = df_census_2021.rename(columns={'LSOA': 'geography code'})
In [21]:
# Use GroupBy() to compute the sum
df2 = df_census_2021.groupby('geography code').sum()
In [22]:
#Joining health data with census data

health_census = pd.merge(left = df2, right = join_data, how = 'left', on = 'geography code' )
In [23]:
health_census.shape
Out[23]:
(312, 31)

2.1 Feature Engineering¶

In [24]:
#Dataset Information for health_census data

print('The Employment Rate dataframe has {} rows and {} columns'.format(health_census.shape[0], health_census.shape[1]))

The Employment Rate dataframe has 312 rows and 31 columns
In [25]:
# Renaming the column in 2021 census data for further joining of data

health_census = health_census.rename(columns={'geography code' : 'LSOA'})
In [26]:
# Number of LAD in UK census dataset

health_census['LSOA'].unique().size
Out[26]:
312
In [27]:
#Description of Datasets

health_census.describe(include='all')
Out[27]:
LSOA Asian Black White Others General health: Total: All usual residents General health: Very good health General health: Good health General health: Fair health General health: Bad health ... Provision of unpaid care: Provides 9 hours or less unpaid care a week Provision of unpaid care: Provides 10 to 19 hours unpaid care a week Provision of unpaid care: Provides 20 to 49 hours unpaid care a week Provision of unpaid care: Provides 20 to 34 hours unpaid care a week Provision of unpaid care: Provides 35 to 49 hours unpaid care a week Provision of unpaid care: Provides 50 or more hours unpaid care a week Number of disabled people in household: Total: All households Number of disabled people in household: No people disabled under the Equality Act in household Number of disabled people in household: 1 person disabled under the Equality Act in household Number of disabled people in household: 2 or more people disabled under the Equality Act in household
count 312 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000 312.00000 312.000000 ... 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000
unique 312 NaN NaN NaN NaN NaN NaN NaN NaN NaN ... NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
top E01010568 NaN NaN NaN NaN NaN NaN NaN NaN NaN ... NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
freq 1 NaN NaN NaN NaN NaN NaN NaN NaN NaN ... NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
mean NaN 562.983974 35.272436 1070.464744 82.612179 1751.272436 808.131410 604.685897 234.38141 79.794872 ... 44.064103 19.666667 37.141026 16.557692 20.583333 45.490385 672.689103 442.785256 180.435897 49.467949
std NaN 615.979284 43.981502 540.291566 52.050497 317.107921 171.413432 113.371586 49.62877 29.266976 ... 18.212811 6.763657 15.465114 7.008713 10.418237 13.535706 136.319735 108.626216 44.958885 14.602744
min NaN 3.000000 0.000000 45.000000 7.000000 1056.000000 409.000000 337.000000 101.00000 19.000000 ... 9.000000 5.000000 3.000000 2.000000 0.000000 6.000000 405.000000 240.000000 87.000000 7.000000
25% NaN 56.000000 8.000000 546.250000 42.000000 1543.000000 691.500000 525.500000 203.75000 58.750000 ... 30.000000 15.000000 25.000000 11.000000 12.000000 37.000000 576.750000 363.000000 147.000000 39.000000
50% NaN 273.500000 19.000000 1202.000000 73.000000 1677.500000 792.000000 587.500000 234.00000 78.000000 ... 41.500000 19.000000 34.500000 16.000000 19.000000 45.000000 654.500000 428.000000 174.500000 48.000000
75% NaN 1054.250000 43.000000 1491.000000 113.250000 1941.750000 910.750000 668.500000 269.25000 103.000000 ... 57.000000 24.000000 46.250000 21.000000 27.250000 54.000000 745.000000 506.000000 202.250000 59.000000
max NaN 2286.000000 408.000000 2361.000000 290.000000 2852.000000 1459.000000 1035.000000 383.00000 178.000000 ... 107.000000 44.000000 87.000000 49.000000 57.000000 99.000000 1107.000000 869.000000 367.000000 97.000000

11 rows × 31 columns

2.1.1 Renaming feature names¶

In [28]:
health_census.rename(columns={'General health: Total: All usual residents': 'General_Health_Total', 
                              'General health: Very good health' : 'Health_Very_Good', 
                              'General health: Good health' : 'Health_Good',
                              'General health: Fair health' : 'Health_Fair', 
                              'General health: Bad health' : 'Health_Bad',
                              'General health: Very bad health' : 'Health_Very_Bad',
                              
                              'Disability: Total: All usual residents' : 'Disability_Total',
                              'Disability: Disabled under the Equality Act' : 'Disabled',
                              'Disability: Disabled under the Equality Act: Day-to-day activities limited a lot' : 'Disability_limited_lot',
                              'Disability: Disabled under the Equality Act: Day-to-day activities limited a little' : 'Disability_limited_little',
                              'Disability: Not disabled under the Equality Act' : 'Disability_No',
                              'Disability: Not disabled under the Equality Act: Has long term physical or mental health condition but day-to-day activities are not limited' : 'Disability_phy_mental_yes',
                              'Disability: Not disabled under the Equality Act: No long term physical or mental health conditions' : 'Disability_phy_mental_no',
                              
                              'Provision of unpaid care: Total: All usual residents aged 5 and over' : 'unpaid_care_total',
                              'Provision of unpaid care: Provides no unpaid care' : 'unpaid_care_no',
                              'Provision of unpaid care: Provides 19 hours or less unpaid care a week' : 'unpaid_care_upto19',
                              'Provision of unpaid care: Provides 9 hours or less unpaid care a week' : 'unpaid_care_upto9',
                              'Provision of unpaid care: Provides 10 to 19 hours unpaid care a week' : 'unpaid_care_10to19',
                              'Provision of unpaid care: Provides 20 to 49 hours unpaid care a week' : 'unpaid_care_20to49',
                              'Provision of unpaid care: Provides 20 to 34 hours unpaid care a week' : 'unpaid_care_20to34',
                              'Provision of unpaid care: Provides 35 to 49 hours unpaid care a week' : 'unpaid_care_35to49',
                              'Provision of unpaid care: Provides 50 or more hours unpaid care a week' : 'unpaid_care_above50',
                              
                              'Number of disabled people in household: Total: All households' : 'household_disabled_total',
                              'Number of disabled people in household: No people disabled under the Equality Act in household' : 'household_disabled_no',
                              'Number of disabled people in household: 1 person disabled under the Equality Act in household' : 'household_disabled_1',
                              'Number of disabled people in household: 2 or more people disabled under the Equality Act in household' : 'household_disabled_above1'
                             }, inplace=True)

health_census.head(2)
Out[28]:
LSOA Asian Black White Others General_Health_Total Health_Very_Good Health_Good Health_Fair Health_Bad ... unpaid_care_upto9 unpaid_care_10to19 unpaid_care_20to49 unpaid_care_20to34 unpaid_care_35to49 unpaid_care_above50 household_disabled_total household_disabled_no household_disabled_1 household_disabled_above1
0 E01010568 42 11 1638 31 1719 822 605 206 61 ... 74 18 26 13 13 31 825 577 207 41
1 E01010569 21 17 1674 42 1756 772 583 257 110 ... 50 19 32 13 19 70 841 498 274 69

2 rows × 31 columns

2.1.2 Reorder the columns¶

In [29]:
#Reordering for columns in the dataset

health_census = health_census.reindex(columns=['LSOA', 'Asian', 'Black', 'White', 'Others',
       'General_Health_Total', 'Health_Very_Good', 'Health_Good',
       'Health_Fair', 'Health_Bad', 'Health_Very_Bad', 'Disability_Total',
       'Disabled', 'Disability_limited_lot', 'Disability_limited_little',
       'Disability_No', 'Disability_phy_mental_yes',
       'Disability_phy_mental_no', 'unpaid_care_total', 'unpaid_care_no',
       'unpaid_care_upto19', 'unpaid_care_upto9', 'unpaid_care_10to19',
       'unpaid_care_20to49', 'unpaid_care_20to34', 'unpaid_care_35to49',
       'unpaid_care_above50', 'household_disabled_total',
       'household_disabled_no', 'household_disabled_1',
       'household_disabled_above1'])

2.1.3 Remove duplicates¶

In [30]:
#identify duplicate rows
duplicateRows = health_census[health_census.duplicated()]

#view duplicate rows
duplicateRows
Out[30]:
LSOA Asian Black White Others General_Health_Total Health_Very_Good Health_Good Health_Fair Health_Bad ... unpaid_care_upto9 unpaid_care_10to19 unpaid_care_20to49 unpaid_care_20to34 unpaid_care_35to49 unpaid_care_above50 household_disabled_total household_disabled_no household_disabled_1 household_disabled_above1

0 rows × 31 columns

In [31]:
print(health_census.duplicated().sum())

0

2.1.4. Imputation : missing values¶

In [32]:
health_census.isnull().sum().sum()
Out[32]:
0

2.1.5 Handling Outliers¶

Detecting Outliers The two most efficient business practices for detecting outliers are: Link: https://www.projectpro.io/article/data-cleaning-techniques/651

Removal: The records containing outliers are removed from the distribution. However, the presence of outliers over multiple variables could result in losing out on a large portion of the datasheet with this method. Replacing values: The outliers could alternatively bed treated as missing values and replaced by using appropriate imputation. Capping: Capping the maximum and minimum values and replacing them with an arbitrary value or a value from a variable distribution. Discretization


5.1-Using Box plots
5.2-Using Scatter plot
5.3-Using Histogram
In [33]:
fig = plt.figure(figsize=(8,4))

<Figure size 800x400 with 0 Axes>

Some outliers represent natural variations in the population, and they should be left as is in your dataset. These are called true outliers.

2.2. Feature scaling : Normalization / Standardization¶

2.2.1 skewness¶

In [34]:
df_num = health_census.select_dtypes(include = ['float64', 'int64'])
df_num.head(2)
Out[34]:
Asian Black White Others General_Health_Total Health_Very_Good Health_Good Health_Fair Health_Bad Health_Very_Bad ... unpaid_care_upto9 unpaid_care_10to19 unpaid_care_20to49 unpaid_care_20to34 unpaid_care_35to49 unpaid_care_above50 household_disabled_total household_disabled_no household_disabled_1 household_disabled_above1
0 42 11 1638 31 1719 822 605 206 61 25 ... 74 18 26 13 13 31 825 577 207 41
1 21 17 1674 42 1756 772 583 257 110 34 ... 50 19 32 13 19 70 841 498 274 69

2 rows × 30 columns

In [35]:
params = {'axes.titlesize':'8', 'xtick.labelsize':'12', 'ytick.labelsize':'12'}
plt.rcParams.update(params)
df_num.drop(columns=[]).hist(figsize=(20, 20), bins=50, xlabelsize=8, ylabelsize=8, ); # ; avoid having the matplotlib verbose informations

2.2.2 Data type conversions¶

In [36]:
health_census.dtypes
Out[36]:
LSOA                         object
Asian                         int64
Black                         int64
White                         int64
Others                        int64
General_Health_Total          int64
Health_Very_Good              int64
Health_Good                   int64
Health_Fair                   int64
Health_Bad                    int64
Health_Very_Bad               int64
Disability_Total              int64
Disabled                      int64
Disability_limited_lot        int64
Disability_limited_little     int64
Disability_No                 int64
Disability_phy_mental_yes     int64
Disability_phy_mental_no      int64
unpaid_care_total             int64
unpaid_care_no                int64
unpaid_care_upto19            int64
unpaid_care_upto9             int64
unpaid_care_10to19            int64
unpaid_care_20to49            int64
unpaid_care_20to34            int64
unpaid_care_35to49            int64
unpaid_care_above50           int64
household_disabled_total      int64
household_disabled_no         int64
household_disabled_1          int64
household_disabled_above1     int64
dtype: object

2.2.3 Near Zero Variance¶

Variance tells us about the spread of the data. It tells us how far the points are from the mean.

In [37]:
np.std(health_census)
Out[37]:
Asian                        614.991346
Black                         43.910962
White                        539.425019
Others                        51.967016
General_Health_Total         316.599327
Health_Very_Good             171.138511
Health_Good                  113.189755
Health_Fair                   49.549173
Health_Bad                    29.220036
Health_Very_Bad               11.213937
Disability_Total             316.598415
Disabled                      66.278355
Disability_limited_lot        43.625236
Disability_limited_little     33.679867
Disability_No                286.706793
Disability_phy_mental_yes     38.642430
Disability_phy_mental_no     290.676994
unpaid_care_total            287.605804
unpaid_care_no               269.812600
unpaid_care_upto19            20.783020
unpaid_care_upto9             18.183600
unpaid_care_10to19             6.752809
unpaid_care_20to49            15.440311
unpaid_care_20to34             6.997472
unpaid_care_35to49            10.401527
unpaid_care_above50           13.513997
household_disabled_total     136.101099
household_disabled_no        108.451995
household_disabled_1          44.886778
household_disabled_above1     14.579323
dtype: float64

2.3. Feature creation¶

In [38]:
data = health_census.copy()
In [39]:
# calculating total population

data ['total_population'] = data['Asian'] + data ['Black'] + data ['White'] + data ['Others']

# Calculating % of people in each ethnic group

data['Asian_pct'] = round((data['Asian'] / data['total_population']) * 100,2)
data['Black_pct'] = round((data['Black'] / data['total_population']) * 100,2)
data['White_pct'] = round((data['White'] / data['total_population']) * 100,2)
data['Others_pct'] = round((data['Others'] / data['total_population']) * 100,2)
In [40]:
# Adding general health

data['Health_Very_Good'] = data['Health_Very_Good'] + data['Health_Good']
data['Health_Very_Bad'] = data['Health_Very_Bad'] + data['Health_Bad']
In [41]:
# Apply percentage for the population data with health indicators

# Calculating total health population as[p] percentage

data['Health_Very_Good_pct'] = round((data['Health_Very_Good'] / data['General_Health_Total']) * 100,2)

data['Health_Fair_pct'] = round((data['Health_Fair'] / data['General_Health_Total']) * 100,2)

data['Health_Very_Bad_pct'] = round((data['Health_Very_Bad'] / data['General_Health_Total']) * 100,2)
In [42]:
# Adding disabilty

data['Disabled_pct'] = round((data['Disabled'] / data['Disability_Total']) * 100,2)
data['Disability_limited_lot_pct'] = round((data['Disability_limited_lot'] / data['Disability_Total']) * 100,2)
data['Disability_limited_little_pct'] = round((data['Disability_limited_little'] / data['Disability_Total']) * 100,2)
data['Disability_No_pct'] = round((data['Disability_No'] / data['Disability_Total']) * 100,2)
data['Disability_phy_mental_yes_pct'] = round((data['Disability_phy_mental_yes'] / data['Disability_Total']) * 100,2)
data['Disability_phy_mental_no_pct'] = round((data['Disability_phy_mental_no'] / data['Disability_Total']) * 100,2)
In [43]:
#unpaid care service

data['unpaid_care_up9to49'] = data['unpaid_care_upto19'] + data['unpaid_care_upto9'] + data['unpaid_care_10to19'] + data['unpaid_care_20to49'] + data['unpaid_care_35to49']
In [44]:
# Adding disabilty

data['unpaid_care_up9to49_pct'] = round((data['unpaid_care_up9to49'] / data['unpaid_care_total']) * 100,2)
data['unpaid_care_no_pct'] = round((data['unpaid_care_no'] / data['unpaid_care_total']) * 100,2)
data['unpaid_care_above50_pct'] = round((data['unpaid_care_above50'] / data['unpaid_care_total']) * 100,2)
In [45]:
data.head()
Out[45]:
LSOA Asian Black White Others General_Health_Total Health_Very_Good Health_Good Health_Fair Health_Bad ... Disabled_pct Disability_limited_lot_pct Disability_limited_little_pct Disability_No_pct Disability_phy_mental_yes_pct Disability_phy_mental_no_pct unpaid_care_up9to49 unpaid_care_up9to49_pct unpaid_care_no_pct unpaid_care_above50_pct
0 E01010568 42 11 1638 31 1719 1427 605 206 61 ... 17.04 6.05 10.99 82.96 8.61 74.35 223 13.74 90.82 1.91
1 E01010569 21 17 1674 42 1756 1355 583 257 110 ... 24.27 11.11 13.16 75.73 6.89 68.83 189 11.37 89.72 4.21
2 E01010570 62 4 1409 22 1497 1259 479 172 58 ... 17.42 7.08 10.35 82.58 9.01 73.56 208 14.50 90.24 2.30
3 E01010571 52 27 1768 96 1944 1588 703 251 84 ... 19.00 7.65 11.35 81.00 7.81 73.19 204 11.27 91.88 2.04
4 E01010572 27 4 1422 29 1483 1239 551 196 35 ... 15.91 4.65 11.26 84.09 8.97 75.12 187 13.19 91.33 1.83

5 rows × 49 columns

In [46]:
# dropping irrelevant columns only required columns

data.drop(['Asian', 'Black', 'White', 'Others', 'total_population',
           'General_Health_Total', 'Health_Very_Good', 'Health_Good',
       'Health_Fair', 'Health_Bad', 'Health_Very_Bad', 'Disability_Total',
       'Disabled', 'Disability_limited_lot', 'Disability_limited_little',
       'Disability_No', 'Disability_phy_mental_yes',
       'Disability_phy_mental_no', 'unpaid_care_total', 'unpaid_care_no',
       'unpaid_care_upto19', 'unpaid_care_upto9', 'unpaid_care_10to19',
       'unpaid_care_20to49', 'unpaid_care_20to34', 'unpaid_care_35to49',
       'unpaid_care_above50', 'household_disabled_total',
       'household_disabled_no', 'household_disabled_1',
       'household_disabled_above1','unpaid_care_up9to49',], axis=1, inplace=True)
In [47]:
data.to_csv('Dataset/Processed_Data/processed_clustering_data.csv')

3. EDA¶

In [48]:
data.describe()
Out[48]:
Asian_pct Black_pct White_pct Others_pct Health_Very_Good_pct Health_Fair_pct Health_Very_Bad_pct Disabled_pct Disability_limited_lot_pct Disability_limited_little_pct Disability_No_pct Disability_phy_mental_yes_pct Disability_phy_mental_no_pct unpaid_care_up9to49_pct unpaid_care_no_pct unpaid_care_above50_pct
count 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000 312.000000
mean 29.667853 1.905609 63.823397 4.602724 80.613269 13.461346 5.926154 17.326090 8.008622 9.317788 82.673910 5.640833 77.032692 11.423750 90.986410 2.801250
std 29.909094 2.151954 32.173490 2.637621 3.760570 2.189483 1.942547 3.448461 2.179853 2.011761 3.448461 2.374730 4.356441 2.150007 1.312167 0.776901
min 0.200000 0.000000 2.200000 0.490000 67.560000 6.230000 1.700000 9.160000 2.600000 4.490000 71.030000 1.370000 66.690000 3.090000 87.990000 0.350000
25% 3.510000 0.487500 29.125000 2.437500 78.517500 12.175000 4.590000 14.800000 6.667500 7.837500 80.857500 3.550000 73.995000 10.040000 90.080000 2.320000
50% 15.905000 1.140000 77.865000 4.185000 80.675000 13.380000 5.705000 17.060000 7.805000 9.290000 82.940000 5.725000 76.560000 11.430000 91.000000 2.730000
75% 59.372500 2.520000 92.470000 5.972500 82.712500 14.770000 7.232500 19.142500 9.270000 10.685000 85.200000 7.662500 80.192500 13.002500 91.750000 3.210000
max 94.010000 14.300000 98.740000 15.930000 92.070000 20.330000 12.510000 28.970000 14.860000 14.110000 90.840000 10.760000 86.560000 16.860000 97.840000 5.730000
In [49]:
#Using a hexbin plot with marginal distributions to distintly assess the spread of good health compared to 
#bad health

#Plotting the hexbin plot for Good Health
sns.set(style="ticks")


x = data['Asian_pct']
y = data['Health_Very_Good_pct']

sns.jointplot(x, y, kind="hex", color="#4CB391")

#Plotting the hexbin box plot for Bad health
sns.set(style="ticks")


x = data['Asian_pct']
y = data['Health_Very_Bad_pct']

sns.jointplot(x, y, kind="hex", color="#4CB391")
Out[49]:
<seaborn.axisgrid.JointGrid at 0x7fa9fc23fc40>
Insight: The hexbin plot shows the relationship between ethnic population percentage in Bradford and the health outcomes. The first plot shows that there is a constant relation between Asian population and the good health population. Same pattern can be seen in the bad health population percentage. However, it is interesting to note that Asians are concentrated in the location where there are good population residents are high.
In [50]:
#Using a hexbin plot with marginal distributions to distintly assess the spread of good health compared to 
#bad health

#Plotting the hexbin plot for Good Health
sns.set(style="ticks")


x = data['White_pct']
y = data['Health_Very_Good_pct']

sns.jointplot(x, y, kind="hex", color="#4CB391")

#Plotting the hexbin box plot for Bad health
sns.set(style="ticks")


x = data['White_pct']
y = data['Health_Very_Bad_pct']

sns.jointplot(x, y, kind="hex", color="#4CB391")
Out[50]:
<seaborn.axisgrid.JointGrid at 0x7fa9ff4cc0d0>
Insight: White ethnic group, when there is high population density, residents have very good as well as very bad health. Therefore, segregation of white ethnic group doesn’t play any role in enhancing the society health.
In [51]:
#Using a hexbin plot with marginal distributions to distintly assess the spread of good health compared to 
#bad health

#Plotting the hexbin plot for Good Health
sns.set(style="ticks")


x = data['Black_pct']
y = data['Health_Very_Good_pct']

sns.jointplot(x, y, kind="hex", color="#4CB391")

#Plotting the hexbin box plot for Bad health
sns.set(style="ticks")


x = data['Black_pct']
y = data['Health_Very_Bad_pct']

sns.jointplot(x, y, kind="hex", color="#4CB391")
Out[51]:
<seaborn.axisgrid.JointGrid at 0x7faa02ec7130>
Insight: Blacks have less population percentage contributing to more good health and less bad health residents.
In [52]:
# Relationship Between Asian and Disabled_pct in Bradford (2021)

figure = px.scatter(data_frame = data, 
                    x="Asian_pct",
                    y="Disabled_pct", 
                    size="Disabled_pct", 
                    trendline="ols", 
                    title = "Relationship Between Asian and Disabled_pct in Bradford (2021)")
figure.show()
Insight: Scatterplot shows the distribution of Asian population with percentage of residents disabled and not disabled. It can be visualized that as percentage of Asians residents increases, percentage of disabled population decreases. Region where there is low measure of Asians seems to have high disabled population. Similarly, when population of Asians increases population with no disability increases.
In [53]:
# Relationship Between Asian and Disability_No_pct in Bradford (2021)

figure = px.scatter(data_frame = data, 
                    x="Asian_pct",
                    y="Disability_No_pct", 
                    size="Disability_No_pct", 
                    trendline="ols", 
                    title = "Relationship Between Asian and Disability_No_pct in Bradford (2021)")
figure.show()
In [54]:
figure = px.scatter(data_frame = data, 
                    x="White_pct",
                    y="Disabled_pct", 
                    size="Disabled_pct", 
                    trendline="ols", 
                    title = "Relationship Between White and Disabled_pct in Bradford (2021)")
figure.show()
In [55]:
# Relationship Between White and Disability_No_pct in Bradford (2021)

figure = px.scatter(data_frame = data, 
                    x="White_pct",
                    y="Disability_No_pct", 
                    size="Disability_No_pct", 
                    trendline="ols", 
                    title = "Relationship Between White and Disability_No_pct in Bradford (2021)")
figure.show()
Insight: Residents with disability and no disability both are found in the neighbourhoods where white people are less in population. However, residents with no disability are more when compared to residents with no disability.
In [56]:
# Relationship Between Black and Disabled_pct in Bradford (2021)

figure = px.scatter(data_frame = data, 
                    x="Black_pct",
                    y="Disabled_pct", 
                    size="Disabled_pct", 
                    trendline="ols", 
                    title = "Relationship Between Black and Disabled_pct in Bradford (2021)")
figure.show()
In [57]:
# Relationship Between Black and Disability_No_pct in Bradford (2021)

figure = px.scatter(data_frame = data, 
                    x="Black_pct",
                    y="Disability_No_pct", 
                    size="Disability_No_pct", 
                    trendline="ols", 
                    title = "Relationship Between Black and Disability_No_pct in Bradford (2021)")
figure.show()
Insight: The above scatterplt displays the relationship between percentage of black populations with disabled and not disabled residents. It can be visualized that both variables are significant change in the attributes.
In [58]:
# Relationship Between Asian_pct and unpaid_care_no_pct in Bradford (2021)

sns.scatterplot(x='Asian_pct', y='unpaid_care_no_pct', data= data, marker = '+')
Out[58]:
<AxesSubplot: xlabel='Asian_pct', ylabel='unpaid_care_no_pct'>
In [59]:
# Relationship Between Asian_pct and unpaid_care_no_pct in Bradford (2021)

sns.scatterplot(x='Asian_pct', y='unpaid_care_above50_pct', data= data, marker = '+')
Out[59]:
<AxesSubplot: xlabel='Asian_pct', ylabel='unpaid_care_above50_pct'>
In [60]:
# Relationship Between Asian_pct and unpaid_care_no_pct in Bradford (2021)

sns.scatterplot(x='Black_pct', y='unpaid_care_no_pct', data= data,  alpha = 0.1)
Out[60]:
<AxesSubplot: xlabel='Black_pct', ylabel='unpaid_care_no_pct'>
In [61]:
# Relationship Between Asian_pct and unpaid_care_no_pct in Bradford (2021)

sns.scatterplot(x='Black_pct', y='unpaid_care_above50_pct', data= data,  alpha = 0.1)
Out[61]:
<AxesSubplot: xlabel='Black_pct', ylabel='unpaid_care_above50_pct'>
In [62]:
# Relationship Between Asian_pct and unpaid_care_no_pct in Bradford (2021)

sns.scatterplot(x='White_pct', y='unpaid_care_no_pct', data= data,  alpha = 0.1)
Out[62]:
<AxesSubplot: xlabel='White_pct', ylabel='unpaid_care_no_pct'>
In [63]:
# Relationship Between Asian_pct and unpaid_care_no_pct in Bradford (2021)

sns.scatterplot(x='White_pct', y='unpaid_care_above50_pct', data= data,  alpha = 0.1)
Out[63]:
<AxesSubplot: xlabel='White_pct', ylabel='unpaid_care_above50_pct'>
In [64]:
fig = px.scatter(data, x='White_pct', y='unpaid_care_no_pct', hover_data=['unpaid_care_up9to49_pct', 'unpaid_care_above50_pct'])
fig.show()
In [65]:
fig = px.scatter(data, x='White_pct', y='unpaid_care_above50_pct', hover_data=['unpaid_care_up9to49_pct', 'unpaid_care_above50_pct'])
fig.show()
Insight: The third attribute in health outcomes is the unpaid care provision. An unpaid care is the help or support provided to people who have long-term physical or mental ill-health conditions and illness or problems. The relation between all the ethnic group and unpaid care provision in Bradford. Two variables were selected here to understand the pattern in the data, namely no unpaid care provision and unpaid care provision more than 50 hours. Overall analysis shows no unpaid care provision is provided to almost 98% of residence.
In [ ]: